vlm_grpo
Collection
VLM with GRPO training for enhanced reasoning
•
2 items
•
Updated
Model Card for Model ID
- Base model: Qwen/Qwen2.5-VL-3B-Instruct
- Training: GRPO with leonardPKU/GEOQA_8K_R1V
- Training log on wandb: https://wandb.ai/ddderek-hk-polyu/easy_r1/runs/d1xtspm0
- Total step of 70, not converged yet
- Downloads last month
- 7
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support